Cluster Analysis of Northern Hemisphere Wintertime 500-hPa Flow Regimes during 1920–2014*

Abstract

Clusters in the Northern Hemisphere wintertime, 10-day low-pass-filtered 500-hPa height field are identified using the method of self-organizing maps (SOMs). Results are based on 1) a 57-winter record of ERA and 2) a 93-winter record of the NOAA Twentieth-Century Reanalysis (20CR). The clusters derived from SOMs appear to be more robust and more linearly independent than their counterparts derived from Ward’s method, and clusters with comparable numbers of member days are more distinctive in terms of the standardized Euclidean distances of their centroids from the centroid of the dataset. The reproducible SOM clusters in the hemispheric domain are 1) the negative polarity of the North Atlantic Oscillation (NAO), 2) a pattern suggestive of Alaska blocking with a downstream wave train extending over North America and the North Atlantic, 3) an enhancement of the climatological-mean stationary wave pattern in the Western Hemisphere that projects positively upon the Pacific–North America (PNA) pattern, and 4) a pattern that projects upon the negative polarity of the PNA pattern. The first three patterns have important impacts on the wintertime climate in North America and Europe. In particular, they are helpful in interpreting prevailing flow patterns during the exceptional winters of 1930–31, 2009–10, and 2013–14. Because of the very limited number of independent samples in a single winter, the number of days per winter in which the circulation resides within individual clusters varies erratically from winter to winter, rendering attribution difficult.

The Circumglobal North American wave pattern and its relation to cold events in eastern North America

“…Bao and Wallace [] recently performed a cluster analysis of Northern Hemisphere 10 day low‐pass filtered 500 hPa geopotential heights. They found four reproducible patterns, three of which they related to the NAO and PNA phases.…”

Classifying the tropospheric precursor patterns of sudden stratospheric warmings

“…We use the self-organizing maps (SOM) method for cluster analysis [e.g. Bao and Wallace, 2015]. In order to assess the distinctiveness of a cluster, we calculate the variance ratio as in Bao and Wallace [2015].…”

“…Bao and Wallace, 2015]. In order to assess the distinctiveness of a cluster, we calculate the variance ratio as in Bao and Wallace [2015]. The variance ratio is obtained by dividing the squared distance between a cluster's centroid and the centroid of the entire data set by the mean squared distance between the individual maps in that cluster and the centroid of the entire data set.…”

Maritime Continent seasonal climate biases in AMIP experiments of the CMIP5 multimodel ensemble

“…Cluster analysis is used for classification of homogeneous climate patterns and weather regimes in climate studies (Unal et al 2003;Bao and Wallace 2015). Apart from that, cluster analysis can be used to group ensemble members for operational forecasting purposes (Molteni et al 1996;Legg et al 2002) and classifying CMIP5 climate change projections (Masson and Knutti 2011;Mizuta et al 2014).…”